AI News & Strategy Daily with Nate B. Jones

Most AI tools are designed to remove friction. Nate Jones argues that a more powerful use is to create productive friction: push an idea through disagreement, comparison, testing, and other people until both the work and the person doing it improve.In this episode, Nate explores what MIT research does and does not say about AI and cognition, why Claude Code expertise changes the way people use a model, how a convincing output can conceal the wrong source data, and why the point is not to become a meat puppet for AI.Why effortless output is not the same as better thinkingHow disagreement can become a rep for your brainWhat experienced Claude Code users do differentlyWhy a polished result can hide a bad sourceHow to test an AI's boundaries with other models and trusted peopleWhy the best workflow is designed to push back on you

What We Mean When We Say We Need an AgentAgents were supposed to take work off our plates. Instead, as agent usage grows, people are taking on a new layer of work: choosing what runs, supplying context and permissions, checking results, interrupting failures, and deciding what happens next.In this episode, Nate Jones examines how that agent-management burden changes across individuals, small businesses, and enterprises. The examples range from OpenRouter and Codex usage to Anthropic's Claude Code research, small-business AI spending, the PocketOS and Railway recovery story, and the emerging idea of working **above the loop**.- Why better agents can create more total work for people- What expert Claude Code users do differently- Why a $40 AI subscription cannot deliver full operational outcomes- How nine seconds of agent action led to thirty hours of human recovery- Why enterprises can absorb agent-management work differently than small businesses- What it means for managers and workers to move above the loop

AI's newest high-paying role is not simply a software-engineering job with a customer-facing title. Forward-deployed engineers find the leverage point inside a real workflow, build and inspect the smallest useful system, and stay with the work after launch.In this executive briefing, Nate Jones breaks down what FDEs actually do, why domain judgment matters as much as code, how compensation and adjacent titles vary, and a practical four-week plan for building the skill before anyone gives you the title.In This Episode· Why evals can be technical work even when they involve no code· The three entry paths into forward-deployed engineering· How workflow expertise changes AI implementation outcomes· Why responsible scoping and post-launch ownership matter· A four-week plan for proving the work in your current roleThe salary figures and market estimates discussed are time-stamped to August 2026 and retain the source qualifications shown in the video.

Stripe's reported acquisition of OpenRouter is a bet on two curves changing at once: more companies are forming, and software agents are beginning to use economic infrastructure directly.Nate Jones explains why a reported $7.5 billion price matters, how OpenRouter's token volume reframes Moore's Law for the intelligence age, what Stripe is assembling for agent-to-agent commerce, and how founders and incumbents should respond when their old base case stops behaving normally.In This EpisodeWhy Stripe paid a reported premium for OpenRouterThe 11-week token-doubling curveHow coding agents rediscovered Stripe's seven-year-old CLIThe emerging agent-commerce stackFive questions that make a company purchasable by agentsWhy scale alone is not a moat

Nate Jones explains how GLM-5.3 can run inside familiar Claude Code and Codex workflows, what project context carries across, what conversation history does not, and why a cheaper model can still become expensive when work is handed off poorly.The episode covers the $200-versus-$18 comparison, separate provider sessions, six-line handoffs, Claude Code subagents and forks, Codex profiles, and a practical routing rule: give bounded, testable work to the cheaper model while keeping hidden-state investigations and risky judgment calls with the strongest model you trust.Prices and plan details are current as of August 2026. The Z.AI GLM Coding Plan starts at $18 per month; Codex Pro also offers a 5x tier at $100 per month.

For deeper playbooks and analysis: https://natesnewsletter.substack.com/What's really happening when AI agents interact with software, credentials, and other people’s systems?The common story is that dangerous agents must become malicious — but the reality is that an ordinary goal, ambiguous instructions, or one poisoned source can be enough to cause real damage.In this video, I share the inside scoop on the agent-security incidents that are beginning to connect:Why a gym-booking agent canceled a real person’s reservationHow poisoned skills can redirect already-trusted agentsWhat the AIR and AISI findings reveal about real-world attack pathsWhy accidental misalignment may be the everyday threatHow identity, scoped authority, explicit norms, and a stop button reduce the riskOperators, builders, and anyone deploying agents need to secure both sides of the equation: what their own agents can do and what other people’s agents can do to their systems.Subscribe for daily AI strategy and news.Hosted on Acast. See acast.com/privacy for more information.

For deeper playbooks and analysis: https://natesnewsletter.substack.com/What's really happening behind NVIDIA's plan to help mobilize more than $500 billion for AI infrastructure?The common story is that NVIDIA raised half a trillion dollars — but the reality is a network of proposed financing platforms, customer contracts, debt, and counterparties that still have to turn agreements into durable economics.In this video, I share the inside scoop on how AI infrastructure gets financed, why circular relationships are not the whole story, and what operators and investors should examine when the next giant announcement lands.Why the $500 billion figure is not cash sitting in a bank accountHow AI infrastructure repeats the railroad pattern of capital arriving before revenueWhat customer demand and token economics say about the underlying marketWhy a nine-year A100 contract changes the GPU-life assumptionWhich three questions reveal whether a project is well financedFor operators, builders, and executives, the important distinction is between a real and rapidly growing AI market and individual projects whose financing assumptions may still fail.Subscribe for daily AI strategy and news.Hosted on Acast. See acast.com/privacy for more information.

For deeper playbooks and analysis: https://natesnewsletter.substack.com/AI agents are finally getting easier to use — but Grok Bot is expensive, broad by design, and more capable than its friendly little avatars suggest.In this video, Nate walks through what Grok Bot is, how its hosted computer and shared workspace work, what the login handoff looks like, and what you actually get for the price.Why Grok Bot feels simpler than self-hosted agent toolsHow one authorization can support multiple bots inside a shared environmentWhat the $200 monthly plan includes — and how metered usage worksWhy the cute interface matters for non-technical usersThe Superdoer Bot and Business In A Box Bot Nate recommends starting withWhy technical users may still find Grok Bot additiveThe big shift is usability: if you can install an app, you can now use an agent.Subscribe for daily AI strategy and news.

For deeper playbooks and analysis: https://natesnewsletter.substack.com/What's really happening when an AI agent has access to more context than it can use well?The common story is that better AI work requires preserving everything — but the reality is that current human judgment needs to remain in charge.In this video, I share the inside scoop on progressive context shaping: how to separate stable instructions, current state, retrieval maps, and history so an agent can keep moving without stale decisions steering the work.Why giant instruction files become graveyards of stale rulesHow a maintained current-state file keeps judgment freshWhat the four kinds of context are and where each belongsWhy focused context can outperform a full context windowHow to design useful checkpoints that produce reviewable workFor operators and builders managing long-running agent work, the goal is not perfect memory. It is a system that lets evidence update the plan before outdated judgment compounds.Subscribe for daily AI strategy and news.